5.8.3. X-ray emission from noncentral cluster galaxies
Recent analyses of the X-ray observation of early-type galaxies made with the Einstein X-ray observatory show that many non-cD ellipticals which are not at the centers of rich clusters are moderately strong X-ray sources (Nulsen et al., 1984; Forman et al., 1985; Trinchieri and Fabbiano 1985; Canizares et al., 1987; Trinchieri et al., 1986). The X-ray luminosities range from Lx 1039 - 1042 erg/s. Many of these X-ray emitting elliptical and S0 galaxies are in the Virgo cluster (Forman et al., 1979; Forman and Jones, 1982; Section 4.5.3). A number of peaks in the X-ray surface brightness of A1367 correspond to the positions of galaxies (Bechtold et al., 1983; Section 4.5.4). There is a strong correlation between the optical and X-ray luminosities of these galaxies, with Lx LB1.5 - 2.0, where LB is the optical (blue) luminosity (Forman et al., 1985; Trinchieri and Fabbiano 1985). The X-ray emission is spatially extended, with a typical maximum radius for the brighter galaxies being Rx 50 kpc, and the X-ray surface brightness is reasonably fit by Ix r-2 in the outer parts, where r is the radius from the galaxy center (Forman et al., 1985).
There are a number of strong arguments, given in the papers just cited, which indicate that the X-rays arise as thermal emission from hot, diffuse gas. In the lower luminosity ellipticals, emission by binary X-ray sources may also play a role. The crude X-ray spectral information available suggests that the gas temperatures are on the order of Tg 107 K (Forman et al., 1985). The observed radial variation of the X-ray surface brightness requires that the gas density vary roughly as g r-3/2. The required gas densities in the inner parts are 0.1 atom / cm3. The total mass of hot gas in these galaxies is roughly Mg 109 - 1011 M (Forman et al., 1985). Assuming a stellar mass-to-light ratio of (M/LB)* 8 M / L, the ratio of gas mass to stellar mass is Mg / M* 0.02.
Previous to the detection of this hot gas, elliptical were believed to be gas-poor systems. It was assumed that the gas ejected by stars in ellipticals was effectively heated by supernovae, forming a strong galactic wind which quickly removed the gas from the galaxy (Mathews and Baker 1971). The X-ray observations show that gas is not being removed rapidly from these galaxies; apparently, the galactic winds do not generally occur, at least at the present time. The expected X-ray luminosities from winds are 10-4 of those observed in elliptical galaxies.
A simple model which fits the properties of these X-ray emitting galaxies is that the bulk of the gas forms a cooling flow, in which gas lost by stars in the galaxy is heated by supernovae, the motions of the gas-ejecting stars, and adiabatic compression, and then cools and slowly flows into the galaxy center (Nulsen et al., 1984; Fabian et al. 1986b; Canizares et al., 1987; Sarazin 1986b, c). There are several simple arguments which support this cooling flow model. First, the amount of gas seen is consistent with the current rates of stellar mass loss in ellipticals. The rate of stellar mass loss is proportional to the stellar density *, so that g = * *. For the stellar population in ellipticals, * is calculated to be * 4.7 × 10-20 s-1 = 1.5 × 10-12 yr-1 (Faber and Gallagher 1976). Thus, over 1010 years, the total mass of gas ejected at the present rate is Mg 0.015M*, which is consistent with the observed gas masses. Second, the cooling times in the gas are quite short. Near the centers of the observed X-ray images, the gas densities and temperatures derived from the observations indicate that tcool 107 yr, and the cooling times typically reach 1010 yr only at the very edge of the observed X-ray images. Thus, any gas introduced into the galaxy will tend to cool. Even if the gas could be heated sufficiently to balance the average cooling rate, gas at temperature 107 K is very thermally unstable (Section 5.7.3), and it would form clumps which would cool rapidly. Third, cooling flow models may lead to a natural explanation for the correlation observed between the X-ray and optical luminosities of elliptical galaxies (Nulsen et al., 1984; Canizares et al., 1987; Sarazin, 1986b, c). Finally, the observed radial variation of the X-ray surface brightness of ellipticals can be understood under the cooling flow hypothesis (Sarazin, 1986b, c).
One important aspect of these models is that the X-ray luminosities and their correlation with the optical luminosities can only be understood if the heating of the gas is primarily due to the motions of the gas-losing stars and to adiabatic compression during the inflow, and not due to supernova heating (Canizares et al., 1987; Sarazin, 1986b, c). This implies that the supernova rates in elliptical galaxies are much smaller than had previously been thought. The low supernova rates would also explain why these galaxies lack galactic winds.
In Section 5.5.5, a method was described which allows the mass in galaxies or clusters to be determined if they contain hot, hydrostatic gas. One very important application of this method would be to measure the masses of elliptical galaxies out to large distances. At present it is unclear whether elliptical galaxies possess extended dark matter. X-ray observations of M87 at the center of the Virgo cluster show that it does have such a halo (Bahcall and Sarazin 1977; Mathews 1978; Fabricant et al., 1980; Fabricant and Gorenstein 1983; Section 5.8.1), but it is unclear whether this mass is associated with the cluster center or with M87. Dark haloes have been deduced for spiral galaxies from the rotational velocities of the neutral hydrogen in the disks of the galaxies far outside of their optically luminous regions. Elliptical galaxies do not possess much neutral hydrogen, and thus this technique cannot be applied to them. The masses of elliptical galaxies can be determined from the orbital velocities of their stars (actually the line-of-sight component of the stellar velocity dispersion; see Dressler, 1979, 1981). Unfortunately, these observations are very difficult and cannot be done in the outermost portions of the galaxies. Moreover, the masses are uncertain because the shapes of the orbits of the stars are not known (they may be radial or isotropic, for example; see Section 2.8). The orbital velocities of globular star clusters can also be used to determine the masses of elliptical galaxies (Hesser et al., 1984). Masses can be derived from studies of binary galaxies, but here the orbital characteristics are even more uncertain (Faber and Gallagher, 1979).
The use of X-ray emitting gas to measure the masses of elliptical galaxies out to large distances has a number of important advantages (Section 5.5.5). For example, this gas can be observed out to very large distances from the center of the galaxy. Also, the orbits of gas particles are known to be isotropic because the gas is a collisional fluid. Thus X-ray measurements can give important information on the possible existence of massive haloes around elliptical galaxies.
Forman et al. (1985) attempted to derive mass profiles by applying the hydrostatic method to the X-ray observations of these normal elliptical galaxies. They concluded that these galaxies do indeed have heavy haloes. However, the errors in the temperature determinations for these galaxies are quite large, and the derived masses are very strongly affected by errors in the temperatures (Section 5.5.5). Trinchieri et al. (1986) find that the temperature errors are so large that the masses cannot be determined with sufficient accuracy to decide whether normal elliptical have missing mass haloes. In any case, this is an ideal problem for a future X-ray observatory, such as AXAF (Chapter 6).
The environment of an early type galaxy may also affect the distribution of its X-ray emitting gas. Intracluster gas might either aid in the retention of gas in a galaxy by providing a confining pressure, or aid in the removal of the gas through ram pressure ablation and other stripping processes (Section 5.9). Of particular interest in this regard are M84 and M86, which are two of the most X-ray luminous ellipticals in the Virgo cluster (Figure 26), and the X-ray emitting galaxies in A1367. In A1367, 11 galaxies were detected in X-rays by Bechtold et al. (Section 4.5.4; Figure 27b), of which 8 were found to be spatially extended. The luminosities of these galaxies range from Lx 1 - 7 × 1041 erg/s. It is possible that pressure confinement plays a role in the X-ray emission from these galaxies (Forman et al., 1979; Fabian et al., 1980; Bechtold et al., 1983). The galaxies detected in A1367 have unusually high X-ray luminosities for their optical luminosities, and do not show the correlation between X-ray and optical luminosities seen in other elliptical galaxies. This suggests that the gas is not gravitationally bound, but rather is confined by the intracluster gas in A1367 (Bechtold et al., 1983). However, the density of the intracluster gas in A1367 is comparable to that in the Virgo cluster, where the galaxies have normal X-ray luminosities. Alternatively, Canizares et al. (1987) have suggested that the clumps of X-ray emission attributed by Bechtold et al. to galaxies are really just fluctuations in the intracluster gas, and are only projected near galaxies by coincidence.
In Virgo, the X-ray emission from M86 is more extended than the emission from M84, and forms a plume extending to the north of the galaxy (Figure 26). Based on their radial velocities, M84 appears to be moving fairly slowly, while M86 is moving rapidly. A galaxy that is moving slowly through the intracluster medium might be able to retain more of the gas produced by stellar mass loss, while a more rapidly moving galaxy would tend to lose its gas due to stripping (Section 5.9; Forman et al., 1979; Fabian et al., 1980; Takeda et al., 1984). Because its velocity is considerably larger than the average in the Virgo cluster, Forman et al. and Fabian et al. suggest that the orbit of M86 will carry it far outside the cluster core, where the intracluster gas density is low. In these outer regions of the cluster, stripping of gas may be ineffective, and the galaxy will accumulate gas (Takeda et al., 1984). Forman et al. and Fabian et al. estimate an orbital period for M86 of roughly 5 × 109 yr. Thus the galaxy could amass roughly 5 × 109 M of gas during each orbit. This gas would be stripped during each passage through the cluster core. In this interpretation, the plume to the north of M86 is the trail of gas being stripped from M86 as it enters the core of the Virgo cluster (Forman et al., 1979; Fabian et al., 1982).